I have data that consist of two kinds of features. How can I select top features from the first kind and top features from the second type and then combine it for the learner and predictor?
For now, I'm using the WEKA node "AttributeSelectedClassifier (3.7)".
I'm not sure what you mean by "two kinds of features". Do you have some numeric columns and some string columns (nominal features)?
Anyway you should give the KNIME educational content a try. Here are some nice examples for feature selection workflows: https://www.knime.com/nodeguide/analytics/preprocessing